#llmtext #llmmadness - 3 Jul 2025 - Hi there, I need some advice on trying to turn some ideas I've been working on about generative AI and aesthetics. I've been working on these ideas, I've got a lot of sort of chats with LLMs and lots of words and words and definitely lots of ideas but I'm just a bit stuck on how to go further. Okay, so at the moment I keep repeating this sort of a structure to the LLMs and it never seems to go anywhere so maybe [[the problem]] is with [[the structure]] but I'll let you know. So I really do want to start, this is [[the structure]] now, I really do want to start with more or less verbatim the introduction that I already have. Okay, especially the early paragraphs those can be pretty much exactly the same. Okay, in particular [[the idea]] that because AI art seems so different to traditional forms of art, as in AI art as in generative output of generative AI seems so different from other sorts of art. We have a choice, one is to say that it's not worthy of [[aesthetic appreciation]] and the other is to say that it is worthy of [[aesthetic appreciation]] but it's a different sort of [[aesthetic appreciation]]. And that is what I am going to be trying to do in this paper. Okay, the second option. I then start by some work by Carlson which is about, the first thing I like that he says is basically to appreciate anything we have to appreciate it for what it in fact is. And secondly that's to do with [[scientific knowledge]] I think. He compares this to somebody knowing about art techniques and art history as having a finer appreciation of artworks than other people. Okay, and he says the same thing about the environment. We have to say well what is it in fact? He says well it's natural and it's an environment, it's not a picture, it's not a pretty ornament, it's this result of [[natural forces]]. And then he says it should, and because it's [[the result]] of [[natural forces]] it should be appreciated in the light of the natural sciences. And he names a lot of natural sciences. He focuses on organicy, biological, even ecological examples. He mentions physics in passing. And I am interested in, I want to give an example at about this point in the paper so it's still quite early on. I want to give an example of somebody who knows a lot about physics being able to appreciate the world around them, or their experience of their world around them, in aesthetically new and interesting ways. Okay. I'll put some bones on this example in a little bit but it should be fairly obvious. So that's the end of the introduction and part one. Now, I want to say next that if we consider LLMs, think of them as simulators. But the simulators apply as regards what [[the author]] Janus calls [[semantic physics]]. As you'll be able to see from the chat, I elaborate on this idea in the paper. And I say, well, the best way of thinking about what token generation is, and this includes both [[the tokens]] generated by the LLM itself, and those tokens inputted by the user in the form of prompts. They're best thought of as [[semantic matter]]. Okay. And again, there'll be details in the copy of the chat to what this amounts to. ***And if, and if you really want to talk about the use of the word of the word, the master, the master, the master, the master, the master, the master, the master, the master, the master, the master. And I'll find it for you again at the great things I've been reading to my mind to see.***